How does a teacher in Indonesia use Gemini to help students who cannot write speak?

A classroom where voice becomes image

Eni Wulandari teaches at SLB Negeri 2 Gunungkidul. Her students range from elementary to senior high level. Each one carries a different disability, a different mental age, a different way of learning. The hardest part, she says, is the weight of individualization. Every student needs a personal learning program. Abstract instructions mean very little without visual support.

She opens Gemini on a Chromebook. She types a prompt. A simple image appears: a child holding a pencil. Another: a child opening a book. Students who cannot read see the picture and know what to do. No translation needed. No intermediary. The visual guide bridges the gap between instruction and understanding.

It works immediately.

When writing is not an option

Some of her students cannot write at all. They speak to Gemini instead. The model listens, then generates an image that matches their spoken idea. A student describes a house with a red roof. The image appears on screen. Another describes a family picnic. The image appears. The teacher calls it a voice for those who cannot write. The students call it their own creation.

The shift is practical. Before, a non-writing student depended on a scribe or a symbol board. Now the student speaks directly to the model. The model returns a visual representation. The student recognizes the result as theirs. The teacher watches the recognition happen in real time.

Building the tool they need

Wulandari and her colleagues are not waiting for a finished product. They are building an AI-powered app that unifies special education planning and programming. The goal is a single system where individual learning plans, visual supports, and voice-driven creation live together. The teacher frames it plainly: AI is not just about saving time. It is about ensuring a student's disability no longer determines how much they get to learn.

Chromebooks run the model. The teacher writes prompts. Students speak them. Images appear. The app is still in development. The classroom already uses the pieces that work.

What the announcement does not say

The blog post gives no numbers. It does not say how many students use the system or how many Chromebooks the school owns. It does not name the specific Gemini model version. It does not describe the technical architecture of the app under development. It does not mention privacy controls, data retention, or offline capability. It does not compare outcomes with previous methods. It offers one teacher's account from one school in Indonesia. The rest is implication.